MAT: A Multimodal Attentive Translator for Image Captioning.
Chang LiuFuchun SunChanghu WangFeng WangAlan L. YuillePublished in: CoRR (2017)
Keyphrases
- image data
- image features
- input image
- image representation
- image classification
- image analysis
- template matching
- image regions
- image content
- low level
- image collections
- multiscale
- test images
- image processing
- single image
- region of interest
- image pixels
- image segmentation
- edge detection
- image retrieval
- similarity measure
- segmentation algorithm
- visual information
- feature points
- lighting conditions
- image set
- pixel values
- salient regions
- segmentation method
- image matching
- vector field
- high resolution
- image structure